addyosmani--agent-skills
Addy Osmani
e6d4005d1c
Three corrections from review feedback: 1. Drop the `name = "..."` line from all 7 TOML files. The official Gemini CLI custom-command schema only defines `prompt` (required) and `description` (optional) — command names are derived from the filename. The extra key was silently ignored but misleading. 2. Strip the `agent-skills:` prefix from skill references in prompt bodies. That prefix is Claude Code plugin-namespace syntax; Gemini CLI skills are referenced by bare name (e.g., `incremental-implementation`, not `agent-skills:incremental-implementation`). 3. Rewrite ship.toml's Phase A invocation paragraph. The previous text referenced Claude Code's `subagent_type` parameter, which doesn't exist in Gemini CLI. Replace with the actual Gemini CLI model: subagents in `.gemini/agents/` are exposed to the main agent as tools of the same name (also callable via `@persona-name`), and the parallel fan-out works by issuing the three tool calls in a single assistant turn. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
16 行
635 B
TOML
16 行
635 B
TOML
description = "Break work into small verifiable tasks with acceptance criteria and dependency ordering"
|
|
|
|
prompt = """
|
|
Invoke the planning-and-task-breakdown skill.
|
|
|
|
Read the existing spec (SPEC.md or equivalent) and the relevant codebase sections. Then:
|
|
|
|
1. Enter plan mode — read only, no code changes
|
|
2. Identify the dependency graph between components
|
|
3. Slice work vertically (one complete path per task, not horizontal layers)
|
|
4. Write tasks with acceptance criteria and verification steps
|
|
5. Add checkpoints between phases
|
|
6. Present the plan for human review
|
|
|
|
Save the plan to tasks/plan.md and task list to tasks/todo.md.
|
|
""" |